Author: Aadharsh Pannirselvam
Compiled by: Plain Language Blockchain
It's simple: chains designed, built, and tuned for applications will shine . Next year's best application chains will be carefully assembled based on primitives and first principles.
The recent influx of developers, users, institutions, and capital into the blockchain differs from previous waves: they possess a specific culture (which can be understood as: the definition of user experience), and they value these cultures more than abstract ideals such as decentralization and censorship resistance. In practice, this sometimes aligns with our existing infrastructure, and sometimes it does not.
For cryptographically abstract, non-technical applications like Blackbird or Farcaster, aspects of user experience that are particularly important, such as centralized design decisions that seemed heretical even three years ago— like colocated nodes, single sequencers, and custom databases—are actually quite reasonable . The same applies to stablecoin chains and exchanges like Hyperliquid* and GTE, which rely on milliseconds, minimum price fluctuations (ticks), and optimal prices.
But this doesn't apply to every new application.
For example, balancing this comfort with centralization is the growing interest in privacy among institutions and retailers . The needs and desired experiences of encrypted applications can be drastically different, and therefore their infrastructure should be as well.
Fortunately, assembling a chain from scratch to fit these specific user experience definitions is far less complex than it was two years ago. Today, it's practically no different from assembling a custom PC .
Of course, you can choose every drive, fan, and cable yourself. But if you don't need that level of granularity (which is likely true), you can use services like Digital Storm or Framework, which offer a range of pre-built custom PCs for different needs. If you're somewhere in between, you can add your own parts to components they've already selected and know will work well together. This gives you greater modularity, flexibility, and the ability to eliminate components you don't actually need , while ensuring the final product performs at a high level.
By assembling and tweaking primitives such as consensus mechanisms, execution layers, data storage, and liquidity , applications create culturally unique forms that continuously reflect diverse needs (in other words, user experience concepts), cater to their unique target audiences, and ultimately retain value. These forms may appear as different as ToughBooks, ThinkPads, desktop tower PCs, or MacBooks, but they also converge and coexist to some extent—not every such computer has its own unique operating system. More importantly, each necessary component becomes a "knob" that the application can iterate and adjust as needed without worrying about disruptive changes to the parent protocol.
Given Circle's acquisition of Malachite, a subsidiary of Informal Systems, sovereignty over custom block space is clearly a broader priority right now . In the coming year, I'm excited to see applications and teams defining and owning their chain resources around primitives and reasonable defaults provided by companies like Commonware and Delta, somewhat like HashiCorp or Stripe Atlas for blockchains and block spaces .
Ultimately, this will enable applications to directly own their cash flow and leverage their unique form of construction to deliver the best user experience in their own way as a lasting moat.
The prediction market will continue to innovate.
One of the most acclaimed applications this week is prediction markets. With weekly trading volume across all crypto venues reaching a record $2 billion , it's clear that this category has taken a meaningful step towards becoming a mainstream consumer product.
This momentum creates a tailwind for adjacent projects aiming to complement or replace current market leaders like Polymarket and Kalshi. But amidst the hype, distinguishing genuine innovation from noise will ultimately be key to deciding what to watch in 2026 .
From a market structure perspective, I am particularly excited about solutions that reduce spreads and deepen open interest . While market creation remains licensed and selective, liquidity in prediction markets remains relatively weak for both manufacturers and takers. There are real opportunities to improve optimal routing systems, different liquidity models, and collateral efficiency through products like lending.
Trading volume categorized by sector is also a major driver of some venues outperforming others. For example, over 90% of Kalshi's trading volume in November came from the sports market , highlighting how some venues are naturally better positioned to compete for favorable liquidity. In contrast, Polymarket's trading volume in crypto-related and political markets is 5 to 10 times that of Kalshi.
Nevertheless, on-chain prediction markets still have a long way to go before achieving true mass adoption. A good reference point is the 2025 Super Bowl; this alone generated $23 billion in trading volume in the off-chain betting market , which is more than 10 times the total daily trading volume of all current on-chain markets.
Closing this gap will require sharp, inspired teams to address core prediction market issues, and I will be closely watching these players in the coming year.
Agentic Curators will expand DeFi
The curatorial layer of DeFi lies at two extremes: purely algorithmic (hard-coded interest rate curves, fixed rebalancing rules) or purely human (risk committees, active managers). Intelligent agent curators represent a third regime: AI agents (LLMs + tools + cycles) that manage curatorial and risk strategies in vaults, lending markets, and structured products. They do more than simply execute fixed rules; they reason about risk, return, and strategy.
Consider the curator role in a Morpho market, where someone must define collateral policies, loan-to-value (LTV) limits, and risk parameters to generate yield products. Today, this is a human bottleneck. Intelligent agents can extend it. Soon, you will see intelligent agent curators competing head-on with algorithmic models and human managers .
When will we see DeFi's "Move 37" (referring to the unexpected brilliant move made by the Go AI AlphaGo against Lee Sedol)?
When I talk to crypto fund managers about AI, I get one of two answers: either LLMs are about to automate every trading desk, or they are “illusion toys” that will never withstand the test of real markets. Both views miss the point of architectural shift. Intelligent agents bring emotionless execution, systemic policy adherence, and flexible reasoning to domains where humans are prone to noise and pure algorithms are too fragile. They are more likely to supervise and/or compose lower-level algorithms than replace them. LLMs act as architects designing a secure shell, while deterministic code remains on the hot-latency path.
When the cost of deep reasoning drops to a few cents, the most profitable vault will not be the one with the smartest humans, but the one with the most computing resources.
Short videos are the new storefront
Short videos are rapidly becoming the default interface for people to discover (and ultimately purchase) content they like. TikTok Shop achieved over $20 billion in gross merchandise volume (GMV) in the first half of 2025, nearly doubling year-over-year, and is quietly training a global audience to view entertainment as a storefront .
In response, Instagram has transformed Reels from a defensive feature into a revenue engine . This format delivers more impressions and accounts for an increasingly larger share of Meta's projected advertising revenue by 2025. Whatnot has proven that real-time, personalized sales conversion rates are unmatched by traditional e-commerce .
The underlying theme is simple: when people watch content in real time, they make decisions faster . Every swipe becomes a decision point. Platforms understand this, which is why the line between the recommendation feed and the checkout process is blurring . The feed is the new point of sale, and every creator is a distribution channel.
AI has further driven this transformation. It has reduced the cost of video production, increased the amount of content , and made it easier for creators and brands to test ideas in real time. More content means a larger conversion surface area, and platforms respond by optimizing every second of videos to drive purchase intent .
Cryptocurrency is perfectly suited to this shift. Faster content demands faster, more cost-effective payment tracks . As shopping becomes frictionless and directly embedded in the content itself, you need a system that can settle micropayments, programmatically allocate and split revenue , and track the impact of chaos on on-chain contributions. Cryptocurrency is built for such processes, and it's hard to imagine a hyperscale streaming-native commerce era without it.
Blockchain will drive the new laws of AI expansion.
Over the past few years, the focus of AI has been on a multi-billion dollar arms race between hyperscale companies and startup giants, while decentralized innovators have been groping in the shadows.
But as attention shifted elsewhere, some crypto-native teams made significant strides in decentralized training and inference , and the frontier of this quiet revolution has slowly moved from the whiteboard to testing and production environments.
Teams like Ritual*, Pluralis, Exo*, Odyn, Ambient, and Bagel are now poised for their prime. This new generation of competitors promises to unleash explosive orthogonal impacts on the foundational trajectory of AI.
By training models in a globally distributed setup, and leveraging novel approaches to asynchronous communication and parallelism validated in production-scale operations, scalability constraints can be broken .
The combination of new consensus mechanisms and privacy primitives makes verifiable and confidential reasoning a very realistic option in the on-chain builder toolkit.
The revolutionary blockchain architecture will combine (true) smart contracts with expressive computing structures, thereby simplifying autonomous AI agents that use cryptocurrency as a medium of exchange .
The groundwork has been completed.
The challenge now is to extend this infrastructure to production environments and demonstrate why blockchain can drive fundamental AI innovations that go beyond philosophical, ideological, or skeuomorphic fundraising experiments .
Real-world assets (RWAs) are poised for real-world adoption.
We've been hearing about tokenization for years, but with the mainstream adoption of stablecoins, the emergence of smooth and robust deposit and withdrawal channels, and clearer regulation and support globally, we're finally seeing large-scale adoption of RWAs . According to RWA.xyz*, as of this writing, issued tokenized assets have exceeded $18 billion , compared to just $3.7 billion a year ago, and I expect this momentum to accelerate in 2026.
It is important to note that tokenization and vaults are different design patterns for RWAs : tokenization creates an on-chain representation of off-chain assets, while vaults create a bridge between on-chain capital and off-chain earnings.
I'm excited to see tokenization and vaults providing access to a wide range of physical and financial assets , from commodities like gold and rare metals to raised credit for working capital and payment financing, to raised and publicly offered equities , and much more global currencies . Let's unleash our imaginations. I'd love to see eggs, GPUs, energy derivatives, earned-wage access, Brazilian government bonds, the Japanese yen , and more, all on-chain!
To be clear, this isn't just about putting more things on-chain. It's about upgrading how the world distributes capital through public blockchains , making opaque, slow, and isolated markets accessible, programmable, and liquid . Once they're on-chain, we'll enjoy the benefits of composability with the DeFi primitives we've already built.
Finally, many of these assets will undoubtedly face challenges in terms of transferability, transparency, liquidity, risk management, and distribution , making the infrastructure to mitigate these challenges equally important and exciting!
A product renaissance driven by intelligent agents is on the horizon.
The next generation of networks will be less influenced by the platforms we scroll through and more by the agents we talk to .
We all know that bots and intelligent agents are contributing rapidly to all network activity . Roughly speaking, they account for about 50% today , including both on-chain and off-chain activities. In the crypto space, bots are increasingly representing us in transactions, curating, assisting, scanning contracts, and taking actions, covering everything from trading tokens and managing vaults to auditing smart contracts and developing games.
This is the age of programmable, agent-driven networks . While we have been in it for some time, 2026 will be the year when crypto product design begins to cater more to robots than humans (in a positive, liberating, and non-dystopian way).
This looks like it's still in development, but I personally hope to spend less time clicking through websites and more time interacting with a simple chat-like interface where I manage on-chain bots . Imagine Telegram, but with specific agents tailored to the application/task. They will be able to formulate and execute complex strategies, search the network for the most relevant information and data , and report on transaction results, risks and opportunities requiring attention, and curated information. I will give them a task, and they will track opportunities, filter out all the noise, and execute at the optimal time .
The infrastructure to achieve this already exists on-chain. By combining the default open data graph and programmatic micropayments with on-chain social graphs and cross-chain liquidity tracks , we have everything needed to support a dynamic ecosystem of intelligent agents. The plug-and-play nature of cryptocurrencies means that agents have fewer red tapes and dead ends to navigate. The readiness of blockchain for this compared to Web2 infrastructure cannot be overstated.
And this might be the most important point here. It's not just about automation; it's about breaking free from Web2 silos, freeing free from friction, and freeing free from waiting. We're seeing this shift happening in search: about 20% of Google searches now generate an AI Overview , and data shows that when people see this overview, they're significantly less likely to click on traditional search result links. Manually sifting through pages is becoming unnecessary. Programmable, agent-driven networks will extend this further to the applications we use , which I think is a good thing.
This era will reduce "doom scrolling." It will reduce panic trading . Time zone differences will be eliminated (no more "waiting for Asia to wake up"). Interacting with the on-chain world will become easier and more expressive for every developer and user.
This cycle will repeat as more assets, systems, and users find ways to get on-chain.
More on-chain opportunities → Deploy more intelligent agents → Unlock more value. Repeat.
But what we build now, and how we build it, will determine whether this agent-driven network becomes merely a layer of noise and automation, or ignites a renaissance of empowering and dynamic products .
